The economic value of fundamental and technical information in emerging currency markets

Abstract

We measure the economic value of information derived from macroeconomic variables and from technical trading rules for emerging markets currency investments. Our analysis is based on a sample of 21 emerging markets with a floating exchange rate regime over the period 1997–2007 and explicitly accounts for trading restrictions on foreign capital movements by using non-deliverable forward data. We document that both types of information can be exploited to implement profitable trading strategies. In line with evidence from surveys of foreign exchange professionals concerning the use of fundamental and technical analysis, we find that combining the two types of information improves the risk-adjusted performance of the investment strategies.

Introduction

The literature on exchange rate forecasting has extensively analyzed the predictive content of two types of information: news on macroeconomic fundamentals as used in structural exchange rate models, and information from historical prices as used in technical trading rules. Meese and Rogoff's (1983) finding that structural models cannot outperform a naive random walk forecast at short horizons still stands after 25 years of intense research, see Cheung et al. (2005) for a recent assessment. There is somewhat more supportive evidence for the usefulness of macroeconomic information for forecasting exchange rates at longer horizons, see Mark, 1995, Kilian, 2001 and Berkowitz and Giorgianni (2001), among others. In general, the performance of technical trading rules at short horizons has been found to be considerably better, see Sweeney, 1986, Levich and Thomas, 1993 and Neely and Weller (1999), with Park and Irwin (2007) and Menkhoff and Taylor (2007) providing recent comprehensive surveys. Nevertheless, Olson, 2004, Pukthuanthong-Le et al., 2007 and Neely et al. (in press) report that the profitability of technical trading rules has weakened substantially in recent years, at least for developed currencies.

The predictive ability of structural exchange rate models and technical trading rules has generally been considered in isolation. This is quite remarkable, in the sense that surveys among foreign exchange market participants invariably indicate that they regard both types of information to be important factors for determining future exchange rate movements, see Taylor and Allen, 1992, Menkhoff, 1997, Lui and Mole, 1998, Cheung and Chinn, 2001, and Gehrig and Menkhoff (2004). Not surprisingly then, most foreign exchange professionals use some combination of fundamental analysis and technical analysis for their own decision making, with the relative weight given to technical analysis becoming smaller as the forecasting (or investment) horizon becomes longer.

The weights assigned to fundamental and technical information for a given horizon may also vary over time. For example, Frankel and Froot (1990) provide empirical evidence for the switch of many professional forecasters from being ‘fundamentalists’ (using structural models and macro data) to acting as ‘chartists’ (using technical trading rules) during the second half of the 1980s. They motivate this changing behavior by the fact that fundamentalists experienced large negative returns in the mid-1980s, when currency prices strongly deviated from their fundamental values. This idea of switching behavior has more recently been formalized in so-called heterogeneous agents models. Brock and Hommes, 1997, Brock and Hommes, 1998 develop equilibrium models in which agents update their beliefs about the future profitability of investment strategies based on their past performance. These models show that rational investors can switch between simple (costless) strategies and sophisticated (costly) strategies. When all investors follow the simple strategy prices may diverge from their fundamental value, making it worthwhile for investors to engage in sophisticated strategies, because expected profits increase. Prices are then pushed back to their fundamental value and the expected net profits for sophisticated investors turn negative. This leads them to switch back to simple and costless strategies that might again result in prices moving away from their fundamental value. These heterogeneous agents models have recently been applied to currency markets, explicitly allowing for the presence of both chartists and fundamentalists, see Chiarella et al. (2006) and De Grauwe and Grimaldi, 2005, De Grauwe and Grimaldi, 2006. The relative importance of these two types of traders (and, hence, the two types of information) varies over time as investors are assumed to switch between strategies according to their relative past performance. De Grauwe and Markiewicz (2006) offer an alternative interpretation of these models, in which market participants combine technical analysis and fundamental information in order to forecast future foreign exchange rates, with weights varying over time as a function of past profitability.

Most research on exchange rate forecasting has focused on developed markets. Scarcely any attention has been paid to emerging market currencies, possibly due to the fact that many emerging countries maintained a fixed or pegged exchange rate regime until fairly recently.¹ Since the mid-1990s, approximately, more and more countries have switched to a floating exchange rate regime. Simultaneously, the emerging currency markets became tradable for currency investors in either the deliverable forward market for currencies without trading restrictions or the non-deliverable forward (NDF) market for currencies with restrictions on foreign capital movements. By now the time series length as well as the cross-sectional breadth are sufficient to warrant a meaningful investigation of exchange rate predictability in emerging markets. To the best of our knowledge we are the first to conduct such an analysis. Previous empirical research on heterogeneous agents models has also been limited to developed currency markets, such as Vigfusson (1997) and De Jong et al. (2006). These studies report only limited empirical evidence supporting the switching behavior between fundamentalist and chartist strategies based on past performance that is assumed in the theoretical models.

In this paper we conduct a comprehensive analysis of the economic value of technical and fundamental information in emerging currency markets. Specifically, we assess the performance of currency trading strategies based on monthly fundamental information derived from the real interest rate differential and GDP growth, as well as technical information in the form of daily moving average trading rules and support and resistance trading rules. We implement these strategies for all freely floating emerging market currencies relative to the US dollar over the period 1997–2007 and use the appropriate historical NDF data for currencies with trading restrictions. We also consider combined strategies in which both chartist and fundamentalist information are used, in line with the actual behavior of market participants, as discussed above. In particular, we examine a dynamic combination scheme with time-varying weights according to the relative profitability of the fundamental and technical strategies. As a benchmark we employ a naive strategy that assigns constant and equal weights to the two types of information. Throughout the empirical analysis, we also consider nine developed currencies as a control sample.

Our results can be summarized as follows. First, both fundamentalist and chartist strategies generate economically and statistically significant Sharpe ratios for emerging currency markets. This finding is consistent with McNown and Wallace (1989), who document that fundamentalist trading strategies perform well in four emerging markets over the period 1972–1986. Our positive results for technical trading rules provide out-of-sample evidence for the profits described by Martin (2001) and Lee et al. (2001a) for the early 1990s.

Second, we document that naively combining chartist and fundamentalist strategies generates positive risk-adjusted returns that are both economically and statistically significant. Moreover, the performance of the combined strategy is much more consistent and stable across currencies than the individual fundamentalist and chartist strategies. This provides convincing evidence for the complementary value of technical and fundamental information as suggested by questionnaires among currency traders. The dynamic combined strategies, where the weights assigned to fundamentalist and chartist strategies vary according to their past performance, increase the profitability of the trading strategy relative to the naive combination only modestly. Thus, we find only limited support for the enhanced profitability of the investment strategies based on the heterogeneous agents models of Chiarella et al. (2006) and De Grauwe and Grimaldi, 2005, De Grauwe and Grimaldi, 2006.

Third, for developed currency markets we find that fundamental trading strategies render statistically and economically significant Sharpe ratios, but this is not the case for the chartist strategies. This result is in line with Abhyankar et al. (2005), who conclude that investors may benefit from fundamental exchange rate models trading the US dollar against the Canadian dollar, Japanese Yen, and British Pound over the period 1977–2000. It also corroborates the findings of Olson, 2004, Pukthuanthong-Le et al., 2007 and Neely et al. (in press), who document that returns to technical trading strategies in developed markets have declined over time. We do find substantial benefits from adding emerging currencies to the developed currency strategies. For either the fundamental, chartist as well as the combined strategy, diversifying across these two types of markets leads to a significantly higher Sharpe ratio compared to the strategies that are limited to developed currencies only.

The remainder of this paper is organized as follows. In Section 2 we describe the data. We examine the performance of the fundamentalist and chartist strategies individually in Sections 3 Fundamentalist trading strategies, 4 Chartist trading strategies, respectively. In Section 5 we integrate the chartist and fundamentalist information into combined strategies. Finally, we conclude in Section 6.